feat: Add UPDATE support

CMakeLists.txt

@@ -9,14 +9,15 @@ set(LOADABLE_EXTENSION_NAME ${TARGET_NAME}_loadable_extension)
 project(${TARGET_NAME})
 include_directories(src/include)
 
-set(EXTENSION_SOURCES src/lance_extension.cpp src/lance_scan.cpp
-                      src/lance_search.cpp src/lance_common.cpp
-                      src/lance_filter_ir.cpp
-                      src/lance_storage.cpp
-                      src/lance_replacement.cpp
-                      src/lance_insert.cpp
-                      src/lance_delete.cpp
-                      src/lance_write.cpp)
+	set(EXTENSION_SOURCES src/lance_extension.cpp src/lance_scan.cpp
+	                      src/lance_search.cpp src/lance_common.cpp
+	                      src/lance_filter_ir.cpp
+	                      src/lance_storage.cpp
+	                      src/lance_replacement.cpp
+	                      src/lance_insert.cpp
+	                      src/lance_delete.cpp
+	                      src/lance_update.cpp
+	                      src/lance_write.cpp)
 
 build_static_extension(${TARGET_NAME} ${EXTENSION_SOURCES})
 build_loadable_extension(${TARGET_NAME} " " ${EXTENSION_SOURCES})

Cargo.toml

@@ -18,12 +18,15 @@ panic = "abort"
 
 [dependencies]
 
-# Lance and Arrow dependencies - using compatible versions
-lance = { version = "1.0.0", default-features = false, features = ["aws", "azure", "gcp", "oss", "huggingface"] }
-lance-core = "1.0.0"
-lance-namespace = "1.0.0"
-lance-namespace-impls = { version = "1.0.0", features = ["rest"] }
-arrow = { version = "56.2.0", features = ["ffi"] }
+	# Lance and Arrow dependencies - using compatible versions
+	lance = { version = "1.0.0", default-features = false, features = ["aws", "azure", "gcp", "oss", "huggingface"] }
+	lance-arrow = "1.0.0"
+	lance-core = "1.0.0"
+	lance-namespace = "1.0.0"
+	lance-namespace-impls = { version = "1.0.0", features = ["rest"] }
+	lance-table = "1.0.0"
+	datafusion = "50.3.0"
+	arrow = { version = "56.2.0", features = ["ffi"] }
 arrow-array = "56.2.0"
 arrow-schema = "56.2.0"
 bzip2-sys = { version = "0.1.13", features = ["static"] }
@@ -36,3 +39,4 @@ datafusion-expr = "50.3.0"
 datafusion-functions = "50.3.0"
 datafusion-sql = "50.3.0"
 lance-index = "1.0.0"
+roaring = "0.10.12"

rust/error.rs

@@ -31,6 +31,7 @@ pub enum ErrorCode {
     DatasetCommitTransaction = 25,
     DirNamespaceDropTable = 26,
     DatasetDelete = 27,
+    DatasetUpdateOverwrite = 28,
 }
 
 struct LastError {

rust/ffi/dataset.rs

@@ -3,9 +3,9 @@ use std::ffi::{c_char, c_void, CStr};
 use std::ptr;
 use std::sync::Arc;
 
+use datafusion_sql::unparser::expr_to_sql;
 use lance::dataset::builder::DatasetBuilder;
 use lance::Dataset;
-use datafusion_sql::unparser::expr_to_sql;
 
 use crate::error::{clear_last_error, set_last_error, ErrorCode};
 use crate::runtime;
@@ -274,7 +274,9 @@ fn dataset_delete_inner(
     };
     let predicate = match filter {
         Some(expr) => expr_to_sql(&expr)
-            .map_err(|err| FfiError::new(ErrorCode::DatasetDelete, format!("predicate sql: {err}")))?
+            .map_err(|err| {
+                FfiError::new(ErrorCode::DatasetDelete, format!("predicate sql: {err}"))
+            })?
             .to_string(),
         None => "true".to_string(),
     };
@@ -315,9 +317,8 @@ fn dataset_delete_inner(
     };
 
     let deleted_rows = before_rows.saturating_sub(after_rows);
-    let deleted_rows_i64 = i64::try_from(deleted_rows).map_err(|_| {
-        FfiError::new(ErrorCode::DatasetDelete, "deleted row count overflow")
-    })?;
+    let deleted_rows_i64 = i64::try_from(deleted_rows)
+        .map_err(|_| FfiError::new(ErrorCode::DatasetDelete, "deleted row count overflow"))?;
 
     unsafe {
         std::ptr::write_unaligned(out_deleted_rows, deleted_rows_i64);

rust/ffi/dir_namespace.rs

@@ -5,8 +5,8 @@ use std::sync::Arc;
 
 use lance::dataset::builder::DatasetBuilder;
 use lance_core::Error as LanceError;
-use lance_namespace::LanceNamespace;
 use lance_namespace::models::{DropTableRequest, ListTablesRequest};
+use lance_namespace::LanceNamespace;
 use lance_namespace_impls::DirectoryNamespaceBuilder;
 
 use crate::error::{clear_last_error, set_last_error, ErrorCode};
@@ -50,10 +50,7 @@ fn parse_storage_options(
             FfiError::new(ErrorCode::Utf8, format!("option_keys[{idx}] utf8: {err}"))
         })?;
         let value = unsafe { CStr::from_ptr(val_ptr) }.to_str().map_err(|err| {
-            FfiError::new(
-                ErrorCode::Utf8,
-                format!("option_values[{idx}] utf8: {err}"),
-            )
+            FfiError::new(ErrorCode::Utf8, format!("option_values[{idx}] utf8: {err}"))
         })?;
         storage_options.insert(key.to_string(), value.to_string());
     }
@@ -123,7 +120,9 @@ fn open_dataset_in_dir_namespace_inner(
     option_values: *const *const c_char,
     options_len: usize,
 ) -> FfiResult<(DatasetHandle, String)> {
-    let root = unsafe { cstr_to_str(root, "root")? }.trim_end_matches('/').to_string();
+    let root = unsafe { cstr_to_str(root, "root")? }
+        .trim_end_matches('/')
+        .to_string();
     let table_name = unsafe { cstr_to_str(table_name, "table_name")? };
     let storage_options = parse_storage_options(option_keys, option_values, options_len)?;
 
@@ -160,7 +159,13 @@ pub unsafe extern "C" fn lance_open_dataset_in_dir_namespace(
         }
     }
 
-    match open_dataset_in_dir_namespace_inner(root, table_name, option_keys, option_values, options_len) {
+    match open_dataset_in_dir_namespace_inner(
+        root,
+        table_name,
+        option_keys,
+        option_values,
+        options_len,
+    ) {
         Ok((handle, table_uri)) => {
             clear_last_error();
             if !out_table_uri.is_null() {
@@ -224,7 +229,8 @@ pub unsafe extern "C" fn lance_dir_namespace_drop_table(
     option_values: *const *const c_char,
     options_len: usize,
 ) -> i32 {
-    match dir_namespace_drop_table_inner(root, table_name, option_keys, option_values, options_len) {
+    match dir_namespace_drop_table_inner(root, table_name, option_keys, option_values, options_len)
+    {
         Ok(()) => {
             clear_last_error();
             0

rust/ffi/mod.rs

@@ -8,5 +8,6 @@ mod scan;
 mod search;
 mod stream;
 mod types;
+mod update;
 mod util;
 mod write;